DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time
Abstract In-situ irradiation transmission electron microscopy (TEM) offers unique insights into the millisecond-timescale post-cascade process, such as the lifetime and thermal stability of defect clusters, vital to the mechanistic understanding of irradiation damage in nuclear materials. Converting...
| Main Authors: | , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2022-09-01
|
| Series: | Scientific Reports |
| Online Access: | https://doi.org/10.1038/s41598-022-19697-1 |
| _version_ | 1828111476618952704 |
|---|---|
| author | Rajat Sainju Wei-Ying Chen Samuel Schaefer Qian Yang Caiwen Ding Meimei Li Yuanyuan Zhu |
| author_facet | Rajat Sainju Wei-Ying Chen Samuel Schaefer Qian Yang Caiwen Ding Meimei Li Yuanyuan Zhu |
| author_sort | Rajat Sainju |
| collection | DOAJ |
| description | Abstract In-situ irradiation transmission electron microscopy (TEM) offers unique insights into the millisecond-timescale post-cascade process, such as the lifetime and thermal stability of defect clusters, vital to the mechanistic understanding of irradiation damage in nuclear materials. Converting in-situ irradiation TEM video data into meaningful information on defect cluster dynamic properties (e.g., lifetime) has become the major technical bottleneck. Here, we present a solution called the DefectTrack, the first dedicated deep learning-based one-shot multi-object tracking (MOT) model capable of tracking cascade-induced defect clusters in in-situ TEM videos in real-time. DefectTrack has achieved a Multi-Object Tracking Accuracy (MOTA) of 66.43% and a Mostly Tracked (MT) of 67.81% on the test set, which are comparable to state-of-the-art MOT algorithms. We discuss the MOT framework, model selection, training, and evaluation strategies for in-situ TEM applications. Further, we compare the DefectTrack with four human experts in quantifying defect cluster lifetime distributions using statistical tests and discuss the relationship between the material science domain metrics and MOT metrics. Our statistical evaluations on the defect lifetime distribution suggest that the DefectTrack outperforms human experts in accuracy and speed. |
| first_indexed | 2024-04-11T11:36:24Z |
| format | Article |
| id | doaj.art-0a4158f87dea43319901cf6a9aad727b |
| institution | Directory Open Access Journal |
| issn | 2045-2322 |
| language | English |
| last_indexed | 2024-04-11T11:36:24Z |
| publishDate | 2022-09-01 |
| publisher | Nature Portfolio |
| record_format | Article |
| series | Scientific Reports |
| spelling | doaj.art-0a4158f87dea43319901cf6a9aad727b2022-12-22T04:25:57ZengNature PortfolioScientific Reports2045-23222022-09-0112111410.1038/s41598-022-19697-1DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-timeRajat Sainju0Wei-Ying Chen1Samuel Schaefer2Qian Yang3Caiwen Ding4Meimei Li5Yuanyuan Zhu6Department of Materials Science and Engineering, University of ConnecticutNuclear Science and Engineering Division, Argonne National LaboratoryDepartment of Materials Science and Engineering, University of ConnecticutDepartment of Computer Science and Engineering, University of ConnecticutDepartment of Computer Science and Engineering, University of ConnecticutNuclear Science and Engineering Division, Argonne National LaboratoryDepartment of Materials Science and Engineering, University of ConnecticutAbstract In-situ irradiation transmission electron microscopy (TEM) offers unique insights into the millisecond-timescale post-cascade process, such as the lifetime and thermal stability of defect clusters, vital to the mechanistic understanding of irradiation damage in nuclear materials. Converting in-situ irradiation TEM video data into meaningful information on defect cluster dynamic properties (e.g., lifetime) has become the major technical bottleneck. Here, we present a solution called the DefectTrack, the first dedicated deep learning-based one-shot multi-object tracking (MOT) model capable of tracking cascade-induced defect clusters in in-situ TEM videos in real-time. DefectTrack has achieved a Multi-Object Tracking Accuracy (MOTA) of 66.43% and a Mostly Tracked (MT) of 67.81% on the test set, which are comparable to state-of-the-art MOT algorithms. We discuss the MOT framework, model selection, training, and evaluation strategies for in-situ TEM applications. Further, we compare the DefectTrack with four human experts in quantifying defect cluster lifetime distributions using statistical tests and discuss the relationship between the material science domain metrics and MOT metrics. Our statistical evaluations on the defect lifetime distribution suggest that the DefectTrack outperforms human experts in accuracy and speed.https://doi.org/10.1038/s41598-022-19697-1 |
| spellingShingle | Rajat Sainju Wei-Ying Chen Samuel Schaefer Qian Yang Caiwen Ding Meimei Li Yuanyuan Zhu DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time Scientific Reports |
| title | DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time |
| title_full | DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time |
| title_fullStr | DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time |
| title_full_unstemmed | DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time |
| title_short | DefectTrack: a deep learning-based multi-object tracking algorithm for quantitative defect analysis of in-situ TEM videos in real-time |
| title_sort | defecttrack a deep learning based multi object tracking algorithm for quantitative defect analysis of in situ tem videos in real time |
| url | https://doi.org/10.1038/s41598-022-19697-1 |
| work_keys_str_mv | AT rajatsainju defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT weiyingchen defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT samuelschaefer defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT qianyang defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT caiwending defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT meimeili defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime AT yuanyuanzhu defecttrackadeeplearningbasedmultiobjecttrackingalgorithmforquantitativedefectanalysisofinsitutemvideosinrealtime |